Probing new physics in diphoton production with proton tagging at the Large Hadron Collider
S. Fichet, G. von Gersdorff, O. Kepka, B. Lenzi, C. Royon, M. Saimpert
aa r X i v : . [ h e p - ph ] D ec Probing new physics in diphoton production with proton tagging at the Large HadronCollider
S. Fichet, ∗ G. von Gersdorff, † O. Kepka, ‡ B. Lenzi, § C. Royon, ¶ and M. Saimpert ∗∗ International Institute of Physics, UFRN, Av. Odilon Gomes de Lima,1722 - Capim Macio - 59078-400 - Natal-RN, Brazil ICTP SAIFR, Instituto de Fisica Teorica, Sao Paulo State University, Brazil Institute of Physics of the Academy of Sciences, Prague CERN, CH-1211 Geneva 23, Switzerland IRFU/Service de Physique des Particules, CEA/Saclay, 91191 Gif-sur-Yvette cedex, France (Dated: December 19, 2013)The sensitivities to anomalous quartic photon couplings at the Large Hadron Collider are esti-mated using diphoton production via photon fusion. The tagging of the protons proves to be a verypowerful tool to suppress the background and unprecedented sensitivities down to 6 · − GeV − are obtained, providing a new window on extra dimensions and strongly-interacting composite statesin the multi-TeV range. Generic contributions to quartic photon couplings from charged and neutralparticles with arbitrary spin are also presented. PACS numbers:
Several major experimental and conceptual facts, likethe overwhelming evidence for dark matter or the gauge-hierarchy problem, point towards the existence of newphysics beyond the Standard Model (SM) at a scale rela-tively close to the electroweak scale. In spite of natural-ness arguments, this pradigm of a TeV-scale new physicsis challenged by both direct searches at the Large HadronCollider (LHC) and by indirect measurements like theLEP electroweak precision tests. In the scenario of newphysics out of reach from direct observation at the LHC,one may expect that the first manifestations show upin precision measurements of the SM properties. Suchpowerful precision tests are already well advanced in theelectroweak and flavour sectors of the SM, and distor-tions of the newly discovered Higgs sector are also beingscrutinized. However, another sector of the SM can betested with high precision at the LHC, the one of puregauge interactions.In this Letter, four-photon (4 γ ) interactions throughdiphoton production via photon fusion with intact out-going protons are considered (Fig. 1). Interactions be- γ γγγpp pp FIG. 1: Diphoton production via photon fusion sensitive to4 γ anomalous couplings. Both protons are intact in the finalstate. tween photons and Z , W bosons in a similar case havealready been studied [1]. The only existing direct lim-its on 4 γ interactions originate from low energy laserexperiments [2]. The study of this process in LHCproton-proton collisions at center-of-mass energy of √ s =14 TeV will benefit from the new forward proton detec-tors considered in the ATLAS and CMS/TOTEM ex-periments [3]. We first provide the generic new physicscontributions to the 4 γ couplings and point out sizablecontributions from strongly-coupled and warped extradimension scenarios. The sensitivities of the upgradesof the ATLAS and CMS/TOTEM experiments are thengiven including all backgrounds.In the assumption of a new physics mass scale Λ heav-ier than experimentally accessible energy E , all newphysics manifestations can be described using an effectiveLagrangian valid for Λ ≫ E . Among these operators, thepure photon dimension-eight operators L γ = ζ γ F µν F µν F ρσ F ρσ + ζ γ F µν F νρ F ρλ F λµ (1)can induce the γγ → γγ process, highly suppressed inthe SM [4, 5]. We discuss here possible new physics con-tributions to ζ γ , that can be probed and discovered atthe LHC using the forward proton detectors.Loops of heavy charged particles contribute to the 4 γ couplings [4] as ζ γi = α Q m − N c i,s , where c ,s = s = 0 − s = − s = 1 , c ,s = s = 0 s = s = 1 (2)where s denotes the spin of the heavy particle of mass m running in the loop and Q its electric charge. The fac-tor N counts all additional multiplicities such as color orflavor. These couplings scale as ∼ Q and are enhancedin presence of particles with large charges. For example,certain light composite fermions, characteristic of com-posite Higgs models, have typically electric charges ofseveral units [4]. For a 500 GeVvector (fermion) reso-nance with Q = 3 (4), large couplings ζ γi of the order of10 − − − GeV − can be reached.Beyond perturbative contributions to ζ γi from chargedparticles, non-renormalizable interactions of neutral par-ticles are also present in common extensions of the SM.Such theories can contain scalar, pseudo-scalar and spin-2 resonances, respectively denoted ϕ , ˜ ϕ , h µν , that coupleto the photon as L γγ = f − + ϕ ( F µν ) + f − − ˜ ϕ F µν F ρλ ǫ µνρλ + f − h µν ( − F µρ F ρν + η µν ( F ρλ ) / , (3)and generate the 4 γ couplings by tree-level exchange as ζ γi = ( f s m ) − d i,s , where d ,s = s = 0 + − s = 0 − − s = 2 , d ,s = s = 0 + s = 0 − s = 2 . (4)Strongly-coupled conformal extensions of the SM con-tain a scalar particle ( s = 0 + ), the dilaton. In the case ofsmall explicit conformal breaking, the dilaton is light andcouples only weakly to the photon, f − ϕ ≪ m − ϕ . How-ever, a more natural situation occurs when explicit con-formal breaking is large [6–8], in which case the dilatonhas a mass comparable to the other resonances of the the-ory and can be much more stongly coupled f − ϕ ∼ π/m ϕ ,as long as photons are mostly composite. In this case,even a 2 TeV dilaton can produce a sizable effective pho-ton interaction, ζ γ ∼ − GeV − .These features are reproduced at large number ofcolours by the gauge-gravity correspondence in a warpedextra dimension. The dilaton is identified as the radion,and a mainly composite photon corresponds to a large in-frared (IR) brane kinetic term. Warped-extra dimensionsalso feature Kaluza-Klein (KK) gravitons [9]. These areinterpreted as spin 2 resonances in the gauge theory. Amostly elementary photon does not yield a sizeable cou-pling. However, a mostly composite one couples morestrongly to the KK fields, in that case the whole set ofKK modes induces [4] ζ γi = κ k d i, , (5)where ˜ k is the IR scale that determines the first KK gravi-ton mass as m = 3 .
83 ˜ k , and κ is a parameter that canbe taken O (1). For κ ∼
1, and m . ζ γ ∼ − GeV − .Since we deal with non-renormalizable couplings per-turbative unitarity (and effective field theory) breaksdown at some scale Λ ′ . This can partially be avoided byusing full amplitudes, but even then some couplings (such as the dilaton coupling f − ϕ ) grow with energy. Wheneverthe scale Λ ′ falls below the detector acceptance a formfactor 1 / (1 + ( m γγ / Λ ′ ) ) is applied to mimic the effectsthat restore unitarity [5]. In many cases such a form fac-tor is not necessary (for instance, when the new particleshave a large enough mass). However, for completenesssensitivities with Λ ′ = 1 TeV are quoted.The γγ → γγ process (Fig. 1) can be probed via thedetection of two intact protons in the forward protondetectors proposed by the ATLAS and CMS [3] collabo-rations, and two energetic photons in the correspondingelectromagnetic calorimeters [10, 11]. The forward detec-tors are expected to be located symmetrically at about210 m from the main interaction point and cover therange 0 . < ξ < .
15, where ξ is the fractional pro-ton momentum loss. The time-of-flight of the scatteredproton can be measured with a precision of ∼
10 ps thatallows to determine the production point of the protonswithin 2.1 mm inside ATLAS/CMS and to check if theyoriginate from the same scattering vertex as the two pho-tons. It is worth noticing that the SM cross section ofdiphoton production with intact protons is dominatedby the QED process at high diphoton mass — and notby gluon exchanges — and is thus very well known. Ifthe protons are not intact, the two-photon quasi-elasticdiphoton production with large theoretical uncertaintiesneeds to be considered [12], leading to a large uncertaintyin the background determination. In the present case,any deviation from the standard model prediction willbe a sign of beyond SM effects.The electromagnetic calorimeters cover the pseudo-rapidity range | η | . . <
1% at transverse momenta p T >
100 GeV) and position (0.001 in η and 1 mradin the azimuthal angle φ ) for photons with p T rangingfrom few GeVto few TeV [13]. A fraction of the photons( ∼ − µ in the following. In thecase of the ATLAS detector, the production point of thephotons can be determined within ∼
15 mm exploitingthe longitudinal segmentation of the ATLAS calorime-ter [14]. Consequently, an alternative scenario with noconverted photons is also considered.According to Ref. [15], even in the presence of morethan 100 pile-up interactions, the photon identifica-tion efficiency is expected to be around 75% for p T > (GeV) γγ m
500 1000 1500 2000 E v en t s -1 SignalExcl. backgroundDPE background,dijet + pile-up - e + e + pile-up γγ = 14 TeVs -1 L = 300 fb = 50 µ -4 GeV -12 = 10 ζ -4 GeV -13 = 10 ζ FIG. 2: Diphoton invariant mass distribution for the signal( ζ = 10 − , − GeV − , see Eq. 1) and for the back-grounds (dominated by γγ with protons from pile-up), re-questing two protons in the forward detectors and two pho-tons of p T >
50 GeV with at least one converted photon in thecentral detector, for a luminosity of 300 fb − and an averagepile-up of µ = 50. Excl. stands for exclusive backgrounds andDPE for double pomeron exchange backgrounds (see text).
100 GeV, with jet rejection factors exceeding 4000. Inaddition, about 1% of the electrons are mis-identified asphotons. These numbers are used in the phenomenolog-ical study presented below.The anomalous γγ → γγ process has been imple-mented in the Forward Physics Monte Carlo (FPMC)generator [16] that aims at providing a variety of diffrac-tive and photon-induced processes in a common frame-work, including a survival probability of 0.9. This fac-tor is necessary to take into account the possibility ofadditional soft interactions occurring between the twointact protons. The FPMC generator was used to sim-ulate the signal and background processes giving rise totwo intact protons accompanied by two photons, elec-trons or jets that can mimic the photon signal. Thoseinclude exclusive SM production of γγ → γγ via leptonand quark boxes and γγ → e + e − . The central exclusiveproduction of γγ via two-gluon exchange, not presentin FPMC, was simulated using ExHuME [17]. This se-ries of backgrounds is called “Exclusive” in Table I andFigs. 2, 3. FPMC was also used to produce γγ , Higgs to γγ and dijet productions via double pomeron exchange(called DPE background in Table I and Fig. 2). Suchbackgrounds tend to be softer than the signal and canbe suppressed with requirements on the transverse mo-menta of the photons ( p T1 >
200 GeV for the leading and p T2 >
100 GeV for the subleading photons, respectively)and the diphoton invariant mass ( m γγ >
600 GeV), asshown in Fig. 2. In addition, the final-state photons ofthe signal are typically back-to-back and have about thesame transverse momenta. Requiring a large azimuthalangle | ∆ φ | > π − .
01 between the two photons and aratio p T2 /p T1 > .
95 greatly reduces the contribution ofnon-exclusive processes.Additional background processes include the quarkand gluon-initiated production of two photons, two jets and Drell-Yan processes leading to two electrons. Thetwo intact protons arise from pile-up interactions (thesebackgrounds are called γγ + pile-up and e + e − , dijet +pile-up in Table I). The hard scattering processes aresimulated with the HERWIG 6.5 [18] generator while thepile-up interactions are simulated by PYTHIA8 [19]. Theprobability to detect at least one proton in each of thetwo forward detectors is estimated to be 32%, 66% and93% for 50, 100 and 200 additional interactions, respec-tively. The pile-up background is further suppressed byrequiring the proton missing invariant mass to match thediphoton invariant mass within the expected resolution( m miss pp = √ ξ ξ s = m γγ ± y pp = 0 . ξ ξ ) to be the same within the resolution( | y γγ − y pp | < . − ( ≃ γγ + pile-up events.Further background reduction is even possible by requir-ing the photons and the protons to originate from thesame vertex that provides an additional rejection factorof 40 for 50 pile-up interactions, showing the large mar-gin on the background suppression. A similar study ata higher pile-up of 200 was performed and led to a neg-ligible background (0.3 expected background events for300 fb − ), showing the robustness of this analysis. More-over, if one relaxes the request of at least one photon tobe converted, the signal is increased by a factor 3 to 4.The sensitivities on photon quartic anomalous couplingsare given in Table II for different scenarios correspondingto the medium luminosity at the LHC (300 fb − ) and thehigh luminosity (6000 fb − when combining the two ex-periments ATLAS and CMS/TOTEM). The sensitivityextends up to 6 · − allowing us to probe further themodels of new physics described above.In this Letter, the sensitivities to quartic photon cou-plings at the LHC, obtained by measuring the photonsin the central CMS and ATLAS detectors and the in-tact protons in dedicated forward proton detectors, areestimated. For the first time, sensitivities on anomalousquartic couplings are large enough to probe models ofnew physics. The imprint of warped KK gravitons andof a strongly-coupled dilaton can be discovered in themulti-TeV range. Also, a generic 500 GeVfermion (vec-tor) resonance can be probed for electric charge Q & TABLE I: Number of signal (for ζ = 2 · − GeV − ) and background events after various selections for an integratedluminosity of 300 fb − and µ = 50 at √ s = 14 TeV. At least one converted photon is required. Excl. stands for exclusivebackgrounds and DPE for double pomeron exchange backgrounds (see text).Cut / Process Signal Excl. DPE e + e − , dijet + pile-up γγ + pile-up0 . < ξ < . p T1 , >
50 GeV 20.8 3.7 48.2 2 . · . · p T1 > p T2 >
100 GeV 17.6 0.2 0.2 1.6 2968 m γγ >
600 GeV 16.6 0.1 0 0.2 1023 p T2 /p T1 > . | ∆ φ | > π − .
01 16.2 0.1 0 0 80.2 √ ξ ξ s = m γγ ±
3% 15.7 0.1 0 0 2.8 | y γγ − y pp | < .
03 15.1 0.1 0 0 0 γγ /m misspp m0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 E v en t s -1 SignalExcl. background + pile-up γγ = 14 TeVs -1 L = 300 fb = 50 µ -4 GeV -12 = 10 ζ -4 GeV -13 = 10 ζ pp -y γγ y-1 -0.5 0 0.5 1 E v en t s -1 SignalExcl. background + pile-up γγ = 14 TeVs -1 L = 300 fb = 50 µ -4 GeV -12 = 10 ζ -4 GeV -13 = 10 ζ FIG. 3: Diphoton to missing proton mass ratio (left) and rapidity difference (right) distributions for signal considering twodifferent coupling values (10 − and 10 − GeV − , see Eq. 1) and for backgrounds after requirements on photon p T , diphotoninvariant mass, p T ratio between the two photons and on the angle between the two photons. At least one converted photonis required. The integrated luminosity is 300 fb − and the average pile-up is µ = 50.TABLE II: 5 σ discovery and 95% CL exclusion limits on ζ and ζ couplings in GeV − (see Eq. 1) with and without formfactor (f.f.), requesting at least one converted photon ( ≥ γ ) or not (all γ ). All sensitivities are given for 300 fb − and µ = 50 pile-up events (medium luminosity LHC) exceptfor the numbers of the last column which are given for 6000fb − and µ = 200 pile-up events (high luminosity LHC).Luminosity 300 fb −
300 fb −
300 fb − − pile-up ( µ ) 50 50 50 200coupling ≥ γ ≥ γ all γ all γ (GeV − ) 5 σ
95% CL 95% CL 95% CL ζ f.f. 1 · − · − · − · − ζ no f.f. 3 · − · − · − · − ζ f.f. 3 · − . · − · − · − ζ no f.f. 7 · − · − · − · − (3) via loop effects. The analysis greatly benefits fromthe kinematical constraints from the photon and protonmeasurements, which allows us to obtain negligible back-grounds.We thank useful discussions with Christophe Grojean. ∗ Electronic address: sylvain.fi[email protected] † Electronic address: gersdorff@gmail.com ‡ Electronic address: [email protected] § Electronic address: [email protected] ¶ Electronic address: [email protected] ∗∗ Electronic address: [email protected][1] E. Chapon, O. Kepka, C. Royon, Phys. Rev.
D81 (2010)074003; O. Kepka and C. Royon, Phys. Rev. D (2008)073005; J. de. Favereau et al., preprint arXiv:0908.2020.[2] M. Bregant et al , Phys. Rev. D (2008) 032006.[3] ATLAS Coll., CERN-LHCC-2011-012; TOTEM Coll.,CERN-LHCC-2013-009.[4] S. Fichet and G. von Gersdorff, preprint arXiv:1311.6815.[5] R. S. Gupta, Phys. Rev. D (2012) 014006.[6] Z. Chacko and R. K. Mishra, Phys. Rev. D (2013)115006.[7] B. Bellazzini, et al., Eur. Phys. J. C (2013) 2333.[8] Z. Chacko, R. K. Mishra and D. Stolarski, JHEP (2013) 121.[9] L. Randall and R. Sundrum, Phys. Rev. Lett. (1999)3370.[10] ATLAS Coll., JINST, Vol. 3 (2008) S08003.[11] CMS Coll., JINST, Vol. 3 (2008) S08004.[12] M. Luszczak, A. Szczurek, preprint arXiv:1309.7201.[13] ATLAS Coll., http://cdsweb.cern.ch/record/1125884,CERN-OPEN-2008-020.[14] ATLAS Coll., Phys. Lett. B (2012) 1.[15] ATLAS Coll., ATL-PHYS-PUB-2013-009.[16] M. Boonekamp et al., preprint arXiv:1102.2531.[17] J. Monk and A. Pilkington, Comput. Phys. Commun. (2006) 232; V.A. Khoze, A.D. Martin, M.G. Ryskin,Eur.Phys.J. C (2008) 363.[18] G. Corcella et al., arXiv:hep-ph/0210213.[19] T. Sjostrand, S. Mrenna and P. Z. Skands, Comput. Phys.Commun.178